AGI: An Alchemist’s Dream
Why Science’s approach to creating Artificial General Intelligence is as hopeless as the Alchemist’s for turning lead into gold.
We have all heard of Artificial Intelligence (AI), whether it is through Sci-Fi films that depict the liberation of robots from their slaveholders or for example on the news when we hear of two AI’s that are shutdown after they start talking to each other in their own language.
But there are two clear distinctions within the AI world, which are between ‘Weak’ AI and ‘Strong’/Artificial General Intelligence (AGI). The former performs cognitive tasks, such as making intelligent searches e.g. finding similarities amongst many images and the latter (which doesn't exist yet) would be able to curate that information in the same way our brains do.
‘Weak’ AI can be compared to a human runner. The best runners are optimised for running and not much else. And so, the idea is that there are many different AI’s performing many different tasks all at the same time. This ‘optimised’ information is then provided to us, the curators, the artists, the project managers, in order to make patterns and turn them into ideas that further the human experience. For instance there are AI’s available right now that can help artists match colours and others that can help compose music. This kind of AI would create a world in which materials are more easily at our fingertips, allowing us to focus on our ideas and our experiences. Sounds promising, doesn’t it?
Nevertheless the engineer’s and scientist’s aren’t content with being curators. Many are on the quest for creating an AGI with all the abilities we have, with the hopes that their creation will surpass The Singularity (the point of no return). They wish to be/create ‘objective’ gods. Gods without ‘errors’, without feelings, either for the purpose of being more amenable slaves or purely because their creators do not understand their own subjective thoughts.
“The alchemist’s dream of making a homunculus in the test tube slowly took the shape of creating robots to work for man” (Ivan Illich, Tools for Conviviality, p.30, 1973)
This fascination with AGI is reminiscent of the era of the ‘Ancient’ Alchemists who attempted to turn lead into gold, create the philosophers stone and Homunculi/Golems who moulded out of clay could replace our slaves. “The illusion prevailed that the machine…could do our labour instead of slaves. it is now time to correct this mistake and shake off the illusion that men are born to be slaveholders and that the only thing wrong in the past was that not all men could be equally so”(Ivan Illich, Tools for Conviviality, p.20, 1973).
But like the Alchemists and pretty much every discipline to date (including science) we have lacked an integrative approach. AGI is meant to replicate our ability to integrate many ideas into one/to explain them simply. However Science has generally preached its truths through objective/reductionist lenses. For instance some scientists within the AI field will compare our thoughts and feelings to ‘machinery’, where they make the assumption that our brains can be repaired or upgraded using the same principles as a car mechanic fixing an engine. But to be integrative, at least according to Ken Wilber, we must consider the subjective, intersubjective, objective and interobjective truths/lenses in order to be holistic in what we do.
For instance it is often argued that the ‘Laws’ of science are comparable to the works of Shakespeare. However “the Second Law of Thermodynamics is nothing more than a working hypothesis… On the other hand — a work by Shakespeare: teeming with the most vital ideas about the inner development of man, showing the whole grandeur and misery of a human existence. How could these two things be equivalent?” (E.F. Schumacher, Small is Beautiful, p.67, 1973) and yet it’s only through the marriage of both works that we could ever hope to understand ourselves and so understand how to create an AGI/Homunculi, let alone fix any of the crises that plague our world.
